Introduction to {targets}

Overview

The {targets} R package offers a superior workflow compared to a script-based approach, addressing several common challenges:

Managing a sequence of computational steps : {targets} allows you to define a pipeline of computational steps, specifying the dependencies and relationships between them.
Selective updates : With {targets}, you only need to update the specific steps affected by a change. It intelligently determines which steps are affected and updates them accordingly, saving time and computational resources.
Ensuring validity of results : {targets} automatically tracks the relationships between steps and ensures that the results at the end of the pipeline remain valid, even if intermediate steps change.

Key advantages of the {targets} workflow over script-based approaches include:

Reproducibility : By explicitly defining dependencies, {targets} ensures that changes in one step trigger the re-evaluation of dependent steps, maintaining reproducibility.
Efficiency : Instead of rerunning all scripts, {targets} selectively executes only the affected steps, optimizing computational efficiency.
Scalability : {targets} provides a clear overview of dependencies, making it easier to manage and scale complex workflows with numerous scripts.
Simplified file management : {targets} automatically manages input and output files, removing the burden of manually tracking file dependencies and their locations.

Open the slides in a new tab here.

Download a new R project as a zip file (green “Code” button -> “Download ZIP”). (If you are familiar with git/github, feel free to clone it directly.)
Examine the file structure, the code, the _targets.R file.
Run tar_make().
Examine everything again and see what’s new.
Put the Poisson regression and table you made earlier into functions and add to the appropriate folders.
Add tar_target() calls to _targets.R.
Add a code chunk to output the table in my-analysis.qmd.
Run tar_outdated().
Run `tar_make().